XCP-ng 8.3 betas and RCs feedback 🚀

gskger

@stormi Great news ! I will free up my playlab as a dedicated testbed. Keep up the good work (and speed to 8.3 final )

gskger

@stormi Did a clean ISO install of XCP-ng 8.3 beta on my three host playlab pool and imported some VMs to play around with. All went well on my aging lab. Next step is to connect to a TrueNAS 13.0-U4 shared storage.

Strebor

Congratulations!

I'm a bit confused about upgrading form 8.3 alpha to 8.3 beta.
If I type yum update on my two pool hosts, there is nothing to update.
Does that mean I'm now running 8.3 beta? Or are there extra steps to take if I want to use the command line upgrade from alfa to beta?

So far, my 2x NUC11TNKi5 testpool is running fine.

olivierlambert

Hi @Strebor

You really should read the blog post in the first message (especially the "Upgrade" section)

Strebor

@olivierlambert Thanks for the quick reply as always!
With respect, I did read that post, and it still leaves me with my questions above.

[TL;DR]
Is 8.3 alpha simply renamed to 8.3 beta, and there's nothing else to do than yum update untill there are no more updates?

or, should I upgrade?

If so, please read my question/remarks below:

[Full question]
I have years of experience of updating/upgrading from Xenserver 5.x all the way to XCP-ng 8.2. But I've never upgraded an alpha version before. I have 8.3-alpha, installed from the alpha2 iso on both hosts.

So, taking your advice, I again look at the section "Alpha testers: update from XCP-ng 8.3 alpha". I click on the "as usual" link, which directs me to updating (not upgrading). Beside the link is the remark (using Xen Orchestra or yum update).

So I go to Upgrade (https://xcp-ng.org/docs/upgrade.html#release-notes-known-issues).

I would have liked to try the yum update (from command line) method , but can't figure out if I have to define the target version, and if so what the name of the tarket version is (export VER=8.3 ?)

Or do you suggest I simply download the iso, dd it to USB stick, reboot boot the master with USB stick, install, reboot and then do the same for the other host server?

Or... is this all not required and am I overthinking this, because yum update was enough?

stormi

@Strebor yum update is enough. If there's nothing to update, this is because you already have an updated alpha, that is, at this stage, beta 1. Alpha, beta, are just milestones on the course of the always evolving 8.3 pre-release, with updated installation images, and dedicated announcements to mark the event.

I actually anticipated your question yesterday and answered it where I thought I'd reach most alpha testers: in the last message on the "XCP-ng 8.3 public alpha " thread.

Strebor

@stormi Thanks! Totally clear to me now.

steff22

Xcp-ng 8.3 beta did not work on Asus pn 52 with RTK8125B nic. Xcp-ng 8.3 does not find any network cards

olivierlambert

Hi!

@steff22 are you suggesting previous versions (eg Alpha) are?

steff22

This post is deleted!

steff22

@olivierlambert

hi yes worked on alpha. Worked on Beta 1 now after I unplugged the power and plugged it back in.

but are the same problem as with alpha, something is wrong with the drivers which make it so that I have 2.5 GB up and 40 mbps down or no speed. with Vm backup restore for me about 4MiB speed with a Samsung 970 PRO with 2 ssd inside Asus PN52.

But after it got internet i see there are 88 patches. can install them and see if there is any change

stormi

@steff22 88 patches on beta1 ? I haven't released any yet. Are you sure they come from our repositories?

steff22

@stormi do not know. But there are 88 patches in Xen orchestra.

did a fresh install from the iso. but suspect that I have chosen the wrong iso in ventoy now. says installer xcp-ng 8.3.0 is this alpha?

steff22

@steff22 was probably alpha that I had to install.

Reinstalled beta 1 now and no patches came up.
but is still the same network problem approx. 4 MiB in speed on backup restore

stormi

@steff22 said in XCP-ng 8.3 beta :

RTK8125B

If I remember what @Andrew and others said correctly, this NIC is hard to support well, despite driver backporting efforts.

steff22

@stormi yes, it seems that way. oh is that such a shame since the Asus pn52 has an AMD Ryzen 9 5900HX 8 cores/16 threads 3.3ghz cpu and up 64 gb ram

Andrew

@steff22 My main home server is an Asus PN63-S1, but that's Intel i7 and i225 ethernet. It's a nice box (for an Intel). I like it for the dual M.2 NVMe so I can mirror my internal data, which has already saved me from Kingston NVMe failure.

I also have an ASRock 4X4 BOX-5600U with the 8125 and it works correctly. I just did a quick NFS read test (using dd) and get 115MBytes/sec (connected to 1G switch). I'll have to move it to the 2.5G switch and see what I get. There have been some reports (in general, not specific to this setup) of issues when using 2.5G connections.

I will build a new driver with the updated code form Realtek and see how it works.

Andrew

@steff22 I tested the stock 8125 driver again on my ASRock AMD system and I have full read and write speed at 1G and 2.5G with native access and VLANs... I'll still try a driver update.

gsrfan01

Not sure if this is related to 8.3 quite yet, but having trouble joining a newly installed 8.3 beta server into a pool that was started from 8.3 alpha and updated.

Attempting to add the new host generates this error POOL_JOINING_HOST_TLS_VERIFICATION_MISMATCH which I only see a single other reference to mentioning CPU settings which as far as I can tell match between the 2 hosts.

steff22

@Andrew now see that I also get 118MBytes/sec (connected to 1G switch.

But I don't have a dedicated 2.5 switch. So it is connected to a 10G switch with rj45 sfp+ module which steps down to 2.5g. This solution worked perfectly in windows 10 with full 2.5G both directions on the same Asus pn52 with CrystalDiskMark over smb I got Sequential reed 0.42 MB/s and Write 292.80 MB/s with Xcp-ng win 10 vm